@Benedict Marco I really appreciate your reply. I got to the site through google and I'm trying it out now. Seems to take quite some time so I will reply here later when my account password hopefully is recovered.
This format is great, Moritz. Please continue to be motivated to make short videos (without basics review) as well as your excellent basics videos. All are welcome!
The assorted components scattered on the wooden background are a great aesthetic touch that keeps your videos from appearing too dry and clinical, but I keep expecting you to pick one up and drop it in the breadboard!
An interesting extension of this method is to vary the DC bias of the input waveform with a potentiometer. By doing so, you can provide a control that sweeps continuously from saw -> tri -> ramp. It's this method that was used on the EMS synthesisers such as the VCS3, Synthi A, Synthi 100 etc., albeit with a typically esoteric single-transistor rectifier. Nobody kept the parts count down like EMS ;)
I actually really enjoyed your consistant use and fairly sophisticated water analogy. But i like the short and more technical format more. Keep up the nice drawings and please never finish that series ;)
I discovery your channel today and i loved it! Amazing! Years ago I joy to use a Rebirth and dream with my own synth! I computer science and electronics is my hobby. Comming soon this is my new project! Thank you! You inspired me to make my own synth! Sorry for my poor english!
Hi Moritz, thanks for this nice video! I have some little tip's for you: please put unused opamps in a stable condition. Simply connect the output of each opamp to the neg. Input and connect the pos. Input to gnd/Center of the power-rails. The spike at the triangle waveform could be smooved if you add a snubber network (R+C in series) from the cathode summingpoint to gnd. Best regards from Kaulsdorf!
I cannot express how happy this video make me. I have never been inspired to get into component level electronics. I am a Reaktor aficionado and Logic Pro user of many years where the environment has a lot of object based processing capabilities for MIDI. This video, although it is the first I have seen of anything you have made, has made me realize that the thought process behind creating analog synths is compatible with my own mental mechanics. Thanks so much for this. You have a new patron, and I look forward to this adventure.
As someone who tends to be appreciated for my own explanations, I have to compliment you on the quality of your explanations in these videos. VERY well done. Anticipates what the learner wants to know, and covers everything lucidly and fully. Kudos!
Another brilliant video!! I'm looking forward to hearing about your insights on voltage control of the resonant filter and eventually a VCA too. For being self-taught, your teaching method is wonderful!! I am also self-taught and love to see and hear other methods and ideas. There is no one-way to do something, but your explanations are very informative and insightful. BTW...I've taken your ideas on digital note control using your V/Oct method and transformed that into using a PIC processor and software. This gives you more control over key scanning and note assignments. The PIC outputs the code as a binary number to send to the D/A for V/Oct control of the oscillators. You now also have a method for a Gate/Trigger output to control an envelope generator. I'm still experimenting, but the results have been promising so far.
Because of finances, i had to sadly enough sell most of my eurorack. This really motivates me to get my empty rack filled up with DIY modules, so you inspired me for my 2021 challenge:) thanks moritz!
I absolutely love these videos, I'm getting a much better understanding of electronics it self from learning about these different tools to the point where i could sometimes kinda guess what the solution might be to a problem that showed up on screen, it's amazing
There is a crack in the wave peak can be eliminated by a 'really big' lowpass filter. But as it is not audible, this is only needed for people listening to their music with an oscilloscope ;D
No, it is quite audible. Despite being very narrow, it has extremely high energy. Yes, while you can filter some high energy out, why not simply make a triangle wave generator that generates a triangle wave? Problems with your proposal are: 1) a narrow spike includes frequencies similar but out of the standard ratios of the audible triangle wave harmonics. A Triangle needs the 3rd harmonic to be 1/9 the amplitude, 5th to be 1/25th, and so on, to sound like a triangle. Throwing in extra energy from the spike will include some energy on the 1st, 3rd, 5th that throws these specific ratios out of proportion. 2) the spike will have 2nd and 4th harmonics, that cannot be filtered out without filtering out the 3rd and 5th, and which have absolutely no place in a triangle wave. 3) the cutoff of this filter must track the oscillator, otherwise it will wipe out the output entirely at high osc frequencies, while leaving a huge amount of spike at low osc frequencies.
@@ian_b > Useful for those of us who suffer from Obsessive Compulsive Waveform Disorder. :) I am not complaining about the waveform as such; if you have a musical application for this sound go for it! My only critique is to pendantically make an ENTIRE VIDEO BASED ON THE CONCEPT THAT THIS IS A TRIANGLE WHEN IT IS NOT EVEN CLOSE. What is the point of using a technical term for something other than its specific meaning? If you like the sound just call it "a converter from Sawtooth to my new 'Moritz' Wave" or something similar? Further, since this video takes the teaching role to people who don't this stuff yet, it's a very strong possibly, perhaps even a certainty, that numerous people getting started in this field will falsely think this is a triangle wave.
I would love more like this! Really nice video. Maybe look into hard sync for one installment? Unless sync is more complicated than I imagine it would be
These videos are really great and i pretty much had to watch all your videos after I encountered the first one. You've given me the weird urge to build my own modular, but I still fear that project is way above my head and I'm musically untalented, so I would not know how to use it anyway. 😅 Two things i'd like to add: At 12:40 you show a schematic that is a wrong solution to the problem. I think it would be good practice to also visually mark these with a big red X or something similar in the top right corner, so people that skip through video without listening notice immediately, that this is not to be used. If I understand this correctly the resulting triangle wave only has half the amplitude of the initial saw wave? I guess that is easily solved by yet another op-amp with a 2:1 resistor ratio. Looking forward to your next videos. 😃
why not start with a breadboard and a few components, just to dip your toes in the water? and yeah, think you're right about making it more visible that that one sheet shows a "wrong" approach. and also about the triangle amplitude!
@@MoritzKlein0 Yea, I think a breadboard would be a good start! Sadly I don't have any components to rely on. So I'd need to get the components first, before I could start experimenting - and I don't really feel like I know what I really need in what amounts, so Its hard for me to find the right balance between "doing lots of small orders and then waiting for parts which is annoying and expensive" and "having a lot of stuff I don't need/want". Sadly I couldn't find something like a "Ben Eater synthesizer kit". 😅 What I found funny: I underestimated the price of potentiometers by a lot. Didn't think 8-bit microcontrollers would be so much cheaper than potentiometers. I feel like these will probably be the most expensive component of a synth made from base components? 🤔 (Please don't read this comment as a "You are now responsible for writing me a shopping list!", that was not my intention, but I couldn't figure out a way to formulate my personal struggles that couldn't be interpreted like that. 😅)
@@Blubb3rbub actually it really depends where you get your pots from (and which make/model). the ones i use are quite reliable and they‘re 1€-1.5€ apiece. but funny you mention ben eater synthesizer kits - i‘ve been considering doing those for some time now. maybe do a test run with, like, 20 kits. only problem is german bureaucracy - need a permit i think (no idea how this works and i‘m afraid i‘ll commit tax fraud involuntarily)
@@MoritzKlein0 I'd definitely be interested in those, I totally understand your concerns, though! My mother works with taxes and self-employment-teaching and recommended registering a Gewerbe for that use-case, which is registered to the tax office using the "Fragebogen zur steuerlichen Erfassung" and doing a Steuererklärung for that business - to be on the safe side. If there are no profits (or less than 410€/year?) there would be no tax to pay, but she says its safer to register it. (Sadly you live too far away to visit her Volkshochschul course on self-employment. 😅) I think having that Gewerbe registered might also be useful for your TH-cam in general, though? It might allow you to tax-deduct stuff you buy for your channel? Not sure at what channel size that is needed/required/useful. Consider all of this hearsay, of course, as this was just the result of a few short App-Messages. 🙂
Hi Daniel. I have the same thing. I have implemented this conversion into my VCO. So I have 4 different wave forms (SAW,SQR,TRI and SINE) but only the Triangle wave has a lower output volume and on my scope its not the same peak to peak as the other three. So I need to op amp this signal again so its matches the other three? How can I do this? I have one op-amp left to use. thanks
Great video! I had been thinking about how to do exactly this recently, and I came to pretty much the same conclusions as you here. The only thing I am planning to do differently is to try getting rid of the sharp "dip" at the peak of the triangle wave with some kind of passive low pass. I don't think finding perfectly matched resistors in the inverter can completely eliminate it, since the "vertical" part of the sawtooth isn't perfect, so even perfectly mirrored ramps will both intersect zero at the same time, leading to their sum being zero at a point.
That spike is caused by 2 slew limitations: in the source signal and in the op-amps of the circuit shown. With TL070 series op amps, it will mostly be outside the audible range.
Three of us just built this and found a few interesting details along the way: - a trimpot in place of one of the 10k resistors makes tuning the peaks to match far, far easier. especially since you can listen to it while adjusting it and don't just have to rely on the oscilloscope or worse, a multimeter - we added a 1nF capacitor across the feedback resistor, and that also helped. This was inspired by seeing a 220pF on René Schmitz's original schematic (though his triangle generator is quite different; maybe in a differential comparator mode?). - on a solderless breadboard, it was easy to get an incredibly accurate triangle. when we moved the circuit to a protoboard, though, the notch at the peak showed up for all of us. I think the solderless breadboard's capacitance contributed somehow, though I need to read up on how to translate that ambient capacitance to an RC circuit... - we're going to try some other TL074s; the market is full of cheap clones, knockoffs and marginally out-of-spec chips and it's hard to be certain, but when you have a few dozen laying around, swapping in other chips until we find a decent one seems like a logical first step. - I also wonder about the possibility of bypassing the first op-amp buffer for one half of the waveform, so that the inverted half is shifted just a touch later relative to the non-inverted half?
Some more experimentation later, and we have a solution we're reasonably happy with: - replace the 10k feedback resistor with a trimpot (or trimpot + resistor) that sweeps a few hundred ohms on either side of 10k - add a 10nF capacitor across that trimpot (so it is a 'feedback capacitor' which by itself would make the op-amp run in integrator mode) The capacitor gets rid of the dip at the peak, but has one minor quirk - it behaves differently at different frequencies. We decided to tune the triangle with the trimpot so that it's perfect around 100hz or so; it becomes less and less perfect at higher frequencies, adding a few more harmonics, but as it does so it also gets less obvious that there are harmonics being introduced. We also did that experimentation on a solderless breadboard and quickly found that a perfboard's results will be slightly different, probably due to stray capacitance from the breadboard. So it may require experimenting with a few different values to see which gives the greatest tuning range without dips; the bigger the cap, the less it will dip but the more pronounced the harmonics when it isn't at the perfectly-tuned frequency. There remains a possibility that an even smaller capacitor, just after the diodes and going to ground, will smooth the peak without quite so much change at higher frequencies. This will smooth both top and bottom edges, though, which will continue to change the character of the triangle.
I'm wondering why you don't get a flat bottom on your output triangle, because of the turn-on voltages of the two diodes? Some fault with the way I'm visualizing this, for sure, but it seems that where the ramp connects with its inverted twin at the bottom, the diodes are not conducting? 🤔
@@MoritzKlein0 I wouldn't pay much attention to "analog" dogma if I was you. I really like your pragmatic from-basic-principles approach. I'll take pragmatic over dogmatic any day! :)
Could maybe use a 20k poti for the voltage divider at the inverter since every potentiometer is a voltage divider on itself. So you could adjust until the triangle fits.
These videos are so qualitative ! Thank you for taking the time to make them. I was wondering if that circuit could work with other waveforms, or at least what it would do to them ?
Like the format definitely, I'm guessing as well you'd be able to pull out the inverted saw (with its own ac coupling) in order to get the option of triangle or inverted saw outputs?
would it be possible to use only the negative mix out of the saw and the ramp for the triangle? then you would get a perfect triangle without messing with the resistor matching.
I'm excited to try this - I've built the VCO up to the point of adding the pitch potentiometer but I dont have a sequencer so I'm looking into writing a simple sequence for my Arduino. If anything I can sample the output and use it in a sampler.
I noticed on the scope trace at the peak of the triangle that there was a line going down to zero. I assumed that is was a kind of crossover distortion caused by the diodes both being in their non conducting region (less that 0.6). Would it be beneficial to use Schottky diodes instead of silicon?
I was wondering, how come the forward voltage drop of the diodes does not cause the cutting off of the square wave when Vout from the inverting buffer is less than the Vforward of the diode?
yeah i skimmed out on the explanation there. it works because the 20k resistor after the two diodes is connected to -12V (instead of ground). so even when the op amps are at 0V output, the diodes stay forward biased.
yeah! if you want to get a 10Vpp output, you could either add another op amp at the end, or you could turn the first buffer into an amplifier with a gain of 2.
correct me if i'm wrong, but i think integrators have to be tuned for the exact frequency of the oscillation. meaning that you can use them to convert a 500 Hz square into a nice triangle, but as soon as that frequency changes, the triangle-shape deteriorates heavily.
@@MoritzKlein0 With what I remember from college, there shouldn't be any weird frequency relations, but there might be a practical reason I'm not aware of
@@ggyshay7687 ah sorry, i was wrong. it's actually the amplitude that decreases (pretty massively) as the input frequency increases. you can try it out here: tinyurl.com/yxgkt3vy
I do not understand how the diode drops disappear? I expected you to do the neg feedback after the top diode and use another op-amp in a +1 configuration to get rid of the bottom diode drop (again with the feedback after the diode) This would also make the group delay between the up and down parts of the wave the same. I have not checked if all that is correct, but my main point is where did the diode drops go? ....or am I missing something?
I'm assuming there's a reason to use this kind of a circuit instead of something more like a bridge rectifier... I just don't know what that reason is. 😆
Another fantastic video. Nice approach to quick and simple tutorials too. If I may ask, where did you learn about all this stuff? You explain it all very well, meaning you have a great knowledge on the subject. Are you self taught? Or did you take lessons / have a mentor? Thanks again.
thanks! i'm self-taught, mostly learned through trial & error on the breadboard. since i don't have a background in EE, i found most tutorials really difficult to follow and so i wondered if there wasn't a more accessible way of teaching this stuff. which lead me to starting this channel!
@@MoritzKlein0 I am glad you started. This is by far the best channel on the subject and you do it well. Keep it up! I am also self taught, dove right into the deep end 4 years ago wanting to clone the TR-808 sounds and now working on a 32-step sequencer without using a microcontroller. It is quite a challenge, since I want it to do all kinds of crazy things, but it is great fun. I will eventually make a groovebox combining a sequencer and all of your ideas. I can imagine something like a Monotribe on steroids haha!
Hi all. may I ask for your help? I have build Moritz Klein's VCO and with success. I added also the saw to tri converter on the same board too and with the wave mixing (blending) option, I can blend between Saw and square to a single output jack. so I had this idea to put a SPDT switch into the circuit and do the same thing with the triangle and sine wave. Now i have this 1 thing that bugs me a bit. I noticed on my oscilloscope that the triangle wave is not as loud as the other three waves. How can I pump it up a bit so it matched the other three? these are nicely 5v.pp. Triangle sits around 2v.pp. thanks in advance for any help or guidance.:)
try this: tinyurl.com/yfcyfe27 - this also fixes the triangle's "buzzyness". be aware that you need to use a precision trimmer to set the waveshape though.
I love these videos! So clear and understandable! Out of curiosity, is there a reason you didn’t put the diodes in a precision rectifier configuration? It’s the same parts, but uses the op-amp to mitigate the diode drop.
mostly because i was just following mfos/ray wilson‘s lead (musicfromouterspace.com/analogsynth_new/SAWTOTRICONVERT/SAWTOTRICONVERT.php) - interesting concept though, hadn’t heard of it yet!
Is the Inverting Buffer considered also Inverting Amplifier with a Gain of 1? The way you call it. Kinda makes more sense calling it a inverting buffer, because it is Designed to only output 1 to 1 Inverted Signal.
Hola... buenisimo video..!!! Hay forma de hacerlo funcionar con fuente simple..??? Si es posible con 9v... No importa si el triangulo no queda perfecto.... Gracias...!!!
Great video, but isn't the triangle wave the half frequency of the sawtooth? So if we switch the sawtooth from our VCO to triangle we suddenly got big drop in the tone of the sound it produces.
I think the frequency is the same. If you look back at the drawing of the overlapping waveforms, the peaks of the input sawtooth wave and output triangle are at the same positions.
Very nice explanation, you are such talented teacher, i've been building the entire VCO serie circuit, it is on sweet spot. One question, why did you use TL074 for this case? Can a TL072 be enough? I'm trying to boarding all in the same pcb but 3xDIP14 chips are quite large. Thanks for the big legacy.
Hi @Moritz Klein and thanks for your instant reply, I really appreciated it ! I've ended just today the layout on a veroboard 21x21 based on TL072 for the SAW-to-TRI, and including the latest of your schematic for PWM and SAW. When I'll end the build I may post something! Till then, thanks a lot for your brilliant work!
Dear Klein, mission accomplished! Your VCO works brilliantly combined with a baby-sequencer, I've just finished the vero-layout and tested it out. It sounds amazing, even if I'm not even closer to synth professional Thanks ! Keep it up!
Hi! I'm in love with your videos, they are super clear and very useful at the same time. However, I have a question about AC coupling. I don't think I really got why we need the resistor going to ground. Especially because I've seen guitar pedals without that
without the resistor, the wave will not be centered around zero volts. think of the resistor like an equalizing valve; i explained this in detail here: th-cam.com/video/QBatvo8bCa4/w-d-xo.html you can check out a simulated example here: tinyurl.com/y42zvbx2 - note how the input sine wave has a 5V DC offset, and the output has the same offset. add a 200k resistor to ground though, and your wave is properly centered: tinyurl.com/y4thaukm i think (correct me if i'm wrong) the reason why some circuits omit the resistor is because they assume the next device will have a resistor to ground at the input terminal.
@@MoritzKlein0 I don't know, that's why I didn't get it... Particularly, I was looking at the schematic here, which doesn't have a resistor to ground at the input: www.wamplerpedals.com/blog/uncategorized/2020/05/how-to-design-a-basic-overdrive-pedal-circuit/ I don't know if either the tone or the volume control solve this, because they are in fact going to ground (especially the volume). Thanks for letting me discover falstad.com, though, I have the feeling I will find it extremely useful
So.. I've been meaning to ask. Do you have any recommendations for op-amp and diode selections to build some of your circuits? Is it the case where you have a couple different types of each and use them for most projects or do you generally end up special ordering for most projects?
he generally put the components used, but if you are having trouble finding some of those, you can always look at the datasheets from said components, and from that data find similar components that can be used as replacements
Hi Moritz! Great videos!!! I love your channel and will be suscribing to your patreon. I built this VCO from scratch and I noticed that, when applying a CV signal, the waveforms at the output increase their gain from 10vpp to 12vpp aprox... is this normal? I've never been aware of this before. Again, kudos on your videos!!! Great work!!! Thank you.
Awesome videos! I've been using them to make my own synths :) I was just curious why you would use an inverting op amp instead of a full bridge rectifier? Is there an advantage?
Can you explain which SAW signal is the input to this circuit? Based on the VCO schematic, is it the TP3 (Test point 3)? Or is it XS4 (The output to the SAW wave jack). It's not clear to me. Love the series, and it has inspired me to get into DIY :D
So far I'm assuming it is TP3. The provided simulation in the description shows the 1uF + 100k which is identical to the TP3 point. I think seeing the 200k value in the video threw me off
Honestly, I find that the reverse is more useful. I have way more use for a saw wave than a triangle. But if you need a triangle wav, use a square wav to feed the integrator. Saw waves are precious. That's why the "super saw" exists and not the super triangle or super square.
@@MoritzKlein0 In an LFO or even say... 1Hz-20Khz? I was pretty sure you can get a stable triangle in the narrow audio frequency from a square. I'm gonna have to do some experiments.
The circuit you need is a Precision Full-wave rectifier; yours is not quite right but it only needs two tiny changes. First, put the lower diode where the feedback resistor is, maintaining the polarity. Then attach the feedback resistor to where the two diodes currently join. Feed the output into a buffer amp as you suggest, but instead of AC coupling it, mix in the original sawtooth as well to level shift the output
A wise person once said “if you can’t make it perfect, make it adjustable”
@Benedict Marco I really appreciate your reply. I got to the site through google and I'm trying it out now.
Seems to take quite some time so I will reply here later when my account password hopefully is recovered.
@Benedict Marco It did the trick and I finally got access to my account again. I'm so happy:D
Thanks so much you saved my account!
@@benedictmarco6242 @Fox Kellan stfu bots
This format is great, Moritz. Please continue to be motivated to make short videos (without basics review) as well as your excellent basics videos. All are welcome!
The assorted components scattered on the wooden background are a great aesthetic touch that keeps your videos from appearing too dry and clinical, but I keep expecting you to pick one up and drop it in the breadboard!
An interesting extension of this method is to vary the DC bias of the input waveform with a potentiometer. By doing so, you can provide a control that sweeps continuously from saw -> tri -> ramp. It's this method that was used on the EMS synthesisers such as the VCS3, Synthi A, Synthi 100 etc., albeit with a typically esoteric single-transistor rectifier. Nobody kept the parts count down like EMS ;)
Please dont stop doing this videos. Your explanations are crystal clear!
I actually really enjoyed your consistant use and fairly sophisticated water analogy. But i like the short and more technical format more.
Keep up the nice drawings and please never finish that series ;)
I discovery your channel today and i loved it! Amazing!
Years ago I joy to use a Rebirth and dream with my own synth!
I computer science and electronics is my hobby. Comming soon this is my new project! Thank you! You inspired me to make my own synth!
Sorry for my poor english!
first viedo that I actually watched. you are really good at explaining. seems to my like lego, love it :)
Hi Moritz, thanks for this nice video! I have some little tip's for you: please put unused opamps in a stable condition. Simply connect the output of each opamp to the neg. Input and connect the pos. Input to gnd/Center of the power-rails. The spike at the triangle waveform could be smooved if you add a snubber network (R+C in series) from the cathode summingpoint to gnd. Best regards from Kaulsdorf!
I cannot express how happy this video make me. I have never been inspired to get into component level electronics. I am a Reaktor aficionado and Logic Pro user of many years where the environment has a lot of object based processing capabilities for MIDI. This video, although it is the first I have seen of anything you have made, has made me realize that the thought process behind creating analog synths is compatible with my own mental mechanics. Thanks so much for this. You have a new patron, and I look forward to this adventure.
glad to hear, much appreciated! yeah i think designing circuits is way more fun and accessible than it seems from a distance!
14:16 soooo beautiful sound!! Thank you so much for your videos!
Honestly, I truly wish you well. I hope you are happy to continue these videos. Really makes me exctied to see new videos from you!
Love the format - I don't feel there's anything missing out of the explanations. Thanks!
As someone who tends to be appreciated for my own explanations, I have to compliment you on the quality of your explanations in these videos. VERY well done. Anticipates what the learner wants to know, and covers everything lucidly and fully. Kudos!
Another brilliant video!! I'm looking forward to hearing about your insights on voltage control of the resonant filter and eventually a VCA too. For being self-taught, your teaching method is wonderful!! I am also self-taught and love to see and hear other methods and ideas. There is no one-way to do something, but your explanations are very informative and insightful.
BTW...I've taken your ideas on digital note control using your V/Oct method and transformed that into using a PIC processor and software. This gives you more control over key scanning and note assignments. The PIC outputs the code as a binary number to send to the D/A for V/Oct control of the oscillators. You now also have a method for a Gate/Trigger output to control an envelope generator. I'm still experimenting, but the results have been promising so far.
I love both formats, and your presentation of the material is top-tier.
Because of finances, i had to sadly enough sell most of my eurorack. This really motivates me to get my empty rack filled up with DIY modules, so you inspired me for my 2021 challenge:) thanks moritz!
I agree, both formats are great! I can’t get enough.
You are so good at teaching electronics! Thank you
This is awesome, yet another high quality video from Mr. Klein. I learn a ton from all these videos, thank you!
Please continue these formats, truly helpful, specially with how detailed your explanations are.
I really like the less basic videos to go a bit more indepth!
This format is great too! You will be the first I'll be supporting on Patreon:) your videos are very helpful and enjoyable.
I absolutely love these videos, I'm getting a much better understanding of electronics it self from learning about these different tools to the point where i could sometimes kinda guess what the solution might be to a problem that showed up on screen, it's amazing
There is a crack in the wave peak can be eliminated by a 'really big' lowpass filter. But as it is not audible, this is only needed for people listening to their music with an oscilloscope ;D
Useful for those of us who suffer from Obsessive Compulsive Waveform Disorder. :)
No, it is quite audible. Despite being very narrow, it has extremely high energy. Yes, while you can filter some high energy out, why not simply make a triangle wave generator that generates a triangle wave? Problems with your proposal are: 1) a narrow spike includes frequencies similar but out of the standard ratios of the audible triangle wave harmonics. A Triangle needs the 3rd harmonic to be 1/9 the amplitude, 5th to be 1/25th, and so on, to sound like a triangle. Throwing in extra energy from the spike will include some energy on the 1st, 3rd, 5th that throws these specific ratios out of proportion. 2) the spike will have 2nd and 4th harmonics, that cannot be filtered out without filtering out the 3rd and 5th, and which have absolutely no place in a triangle wave. 3) the cutoff of this filter must track the oscillator, otherwise it will wipe out the output entirely at high osc frequencies, while leaving a huge amount of spike at low osc frequencies.
@@ian_b > Useful for those of us who suffer from Obsessive Compulsive Waveform Disorder. :)
I am not complaining about the waveform as such; if you have a musical application for this sound go for it! My only critique is to pendantically make an ENTIRE VIDEO BASED ON THE CONCEPT THAT THIS IS A TRIANGLE WHEN IT IS NOT EVEN CLOSE. What is the point of using a technical term for something other than its specific meaning? If you like the sound just call it "a converter from Sawtooth to my new 'Moritz' Wave" or something similar? Further, since this video takes the teaching role to people who don't this stuff yet, it's a very strong possibly, perhaps even a certainty, that numerous people getting started in this field will falsely think this is a triangle wave.
It is *very* audible.
Didn't understand a thing, but it sounded jolly fascinating!
That was very well explained and presented. It will be useful for many people for many years.
Very good video ! Clear explanation ! I would love to see more videos like this . Thank you .
I would love more like this! Really nice video. Maybe look into hard sync for one installment? Unless sync is more complicated than I imagine it would be
Thanks! I really liked the level of detail and the length of this video.
Man! I have been loving these. I just started my VCO project with great success. Thanks so much. Keep up the good work!!
enjoyed this format
my best teacher on yt
Thank you for these videos
loved this video format
These videos are really great and i pretty much had to watch all your videos after I encountered the first one. You've given me the weird urge to build my own modular, but I still fear that project is way above my head and I'm musically untalented, so I would not know how to use it anyway. 😅
Two things i'd like to add: At 12:40 you show a schematic that is a wrong solution to the problem. I think it would be good practice to also visually mark these with a big red X or something similar in the top right corner, so people that skip through video without listening notice immediately, that this is not to be used.
If I understand this correctly the resulting triangle wave only has half the amplitude of the initial saw wave? I guess that is easily solved by yet another op-amp with a 2:1 resistor ratio.
Looking forward to your next videos. 😃
why not start with a breadboard and a few components, just to dip your toes in the water? and yeah, think you're right about making it more visible that that one sheet shows a "wrong" approach. and also about the triangle amplitude!
@@MoritzKlein0 Yea, I think a breadboard would be a good start! Sadly I don't have any components to rely on. So I'd need to get the components first, before I could start experimenting - and I don't really feel like I know what I really need in what amounts, so Its hard for me to find the right balance between "doing lots of small orders and then waiting for parts which is annoying and expensive" and "having a lot of stuff I don't need/want". Sadly I couldn't find something like a "Ben Eater synthesizer kit". 😅
What I found funny: I underestimated the price of potentiometers by a lot. Didn't think 8-bit microcontrollers would be so much cheaper than potentiometers. I feel like these will probably be the most expensive component of a synth made from base components? 🤔
(Please don't read this comment as a "You are now responsible for writing me a shopping list!", that was not my intention, but I couldn't figure out a way to formulate my personal struggles that couldn't be interpreted like that. 😅)
@@Blubb3rbub actually it really depends where you get your pots from (and which make/model). the ones i use are quite reliable and they‘re 1€-1.5€ apiece. but funny you mention ben eater synthesizer kits - i‘ve been considering doing those for some time now. maybe do a test run with, like, 20 kits. only problem is german bureaucracy - need a permit i think (no idea how this works and i‘m afraid i‘ll commit tax fraud involuntarily)
@@MoritzKlein0 I'd definitely be interested in those, I totally understand your concerns, though! My mother works with taxes and self-employment-teaching and recommended registering a Gewerbe for that use-case, which is registered to the tax office using the "Fragebogen zur steuerlichen Erfassung" and doing a Steuererklärung for that business - to be on the safe side. If there are no profits (or less than 410€/year?) there would be no tax to pay, but she says its safer to register it. (Sadly you live too far away to visit her Volkshochschul course on self-employment. 😅)
I think having that Gewerbe registered might also be useful for your TH-cam in general, though? It might allow you to tax-deduct stuff you buy for your channel? Not sure at what channel size that is needed/required/useful. Consider all of this hearsay, of course, as this was just the result of a few short App-Messages. 🙂
Oooooh. This gives me ideas for various waveshaping thingymajoscillators. Thanks!
Given that the resultant triangle is half the amplitude of the input sawtooth, I am guessing you could use the last op amp in the chip to adjust that.
yes! (though there are 2 op amps left, not one!)
Hi Daniel. I have the same thing. I have implemented this conversion into my VCO. So I have 4 different wave forms (SAW,SQR,TRI and SINE) but only the Triangle wave has a lower output volume and on my scope its not the same peak to peak as the other three. So I need to op amp this signal again so its matches the other three? How can I do this? I have one op-amp left to use. thanks
Great video! I had been thinking about how to do exactly this recently, and I came to pretty much the same conclusions as you here. The only thing I am planning to do differently is to try getting rid of the sharp "dip" at the peak of the triangle wave with some kind of passive low pass. I don't think finding perfectly matched resistors in the inverter can completely eliminate it, since the "vertical" part of the sawtooth isn't perfect, so even perfectly mirrored ramps will both intersect zero at the same time, leading to their sum being zero at a point.
yeah true. i think i’ll do the same.
That spike is caused by 2 slew limitations: in the source signal and in the op-amps of the circuit shown. With TL070 series op amps, it will mostly be outside the audible range.
Three of us just built this and found a few interesting details along the way:
- a trimpot in place of one of the 10k resistors makes tuning the peaks to match far, far easier. especially since you can listen to it while adjusting it and don't just have to rely on the oscilloscope or worse, a multimeter
- we added a 1nF capacitor across the feedback resistor, and that also helped. This was inspired by seeing a 220pF on René Schmitz's original schematic (though his triangle generator is quite different; maybe in a differential comparator mode?).
- on a solderless breadboard, it was easy to get an incredibly accurate triangle. when we moved the circuit to a protoboard, though, the notch at the peak showed up for all of us. I think the solderless breadboard's capacitance contributed somehow, though I need to read up on how to translate that ambient capacitance to an RC circuit...
- we're going to try some other TL074s; the market is full of cheap clones, knockoffs and marginally out-of-spec chips and it's hard to be certain, but when you have a few dozen laying around, swapping in other chips until we find a decent one seems like a logical first step.
- I also wonder about the possibility of bypassing the first op-amp buffer for one half of the waveform, so that the inverted half is shifted just a touch later relative to the non-inverted half?
Some more experimentation later, and we have a solution we're reasonably happy with:
- replace the 10k feedback resistor with a trimpot (or trimpot + resistor) that sweeps a few hundred ohms on either side of 10k
- add a 10nF capacitor across that trimpot (so it is a 'feedback capacitor' which by itself would make the op-amp run in integrator mode)
The capacitor gets rid of the dip at the peak, but has one minor quirk - it behaves differently at different frequencies. We decided to tune the triangle with the trimpot so that it's perfect around 100hz or so; it becomes less and less perfect at higher frequencies, adding a few more harmonics, but as it does so it also gets less obvious that there are harmonics being introduced.
We also did that experimentation on a solderless breadboard and quickly found that a perfboard's results will be slightly different, probably due to stray capacitance from the breadboard. So it may require experimenting with a few different values to see which gives the greatest tuning range without dips; the bigger the cap, the less it will dip but the more pronounced the harmonics when it isn't at the perfectly-tuned frequency.
There remains a possibility that an even smaller capacitor, just after the diodes and going to ground, will smooth the peak without quite so much change at higher frequencies. This will smooth both top and bottom edges, though, which will continue to change the character of the triangle.
Your videos are brilliant, thank you so much!!
Good level of detail!
More great fundamental explanations, thanks!
I stress to anyone watching or reading this. There iS real art into making precision resistors. They are fun to make and easy to match.
ich mag den style deiner Videos! Optik und Klang is echt schön :)
You are an amazing teacher :)
I'm wondering why you don't get a flat bottom on your output triangle, because of the turn-on voltages of the two diodes? Some fault with the way I'm visualizing this, for sure, but it seems that where the ramp connects with its inverted twin at the bottom, the diodes are not conducting? 🤔
yes, i cut corners explaining things there. the 20k resistor to -12V after the two diodes will ensure that those diodes are always forward biased!
@@MoritzKlein0 Ah yes of course.. : )
Great Scott, you got it Jack Ruby!
Love these videos. Very well explained.
This is great! Maybe later you could follow up with a triangle core design and explain the pros and cons?
think i‘ll do that, but with op amps instead of the 40106 (people have been complaining that the 40106 is not „analog“ - interesting)
@@MoritzKlein0 I wouldn't pay much attention to "analog" dogma if I was you. I really like your pragmatic from-basic-principles approach. I'll take pragmatic over dogmatic any day! :)
@@MoritzKlein0 wait, why are people complaining? it simply generates the waves just like its supposed to do and its not a digital chip
If you offset the sawtooth wave like you showed at 10:55, would it be possible to splice it into a sharktooth wave?
Keep going! I love your explanations!
Wow that sequence!! So goood🎉
Could maybe use a 20k poti for the voltage divider at the inverter since every potentiometer is a voltage divider on itself. So you could adjust until the triangle fits.
a regular potentiometer would a pain to fine-tune though, i think you'd have to go with a precision (10-turn) trimmer!
@@MoritzKlein0 No, just put a 1K at the summing node. As long as its value covers the potential mismatch.
The triangle waves have a pulse wave in the middle of them, producing a strong buzz. How would you get rid of that?
These videos are so qualitative ! Thank you for taking the time to make them.
I was wondering if that circuit could work with other waveforms, or at least what it would do to them ?
I appreciate your work a lot!
Like the format definitely, I'm guessing as well you'd be able to pull out the inverted saw (with its own ac coupling) in order to get the option of triangle or inverted saw outputs?
yes definitely!
@@MoritzKlein0 excellent, thank you!
Why not using a 20k trim pot with the central point connected to the inverting pin of the opamp and the leads as two 10k resistors?
that should work i guess!
Like the short format without the basics!
Thanks for the great work.
would it be possible to use only the negative mix out of the saw and the ramp for the triangle? then you would get a perfect triangle without messing with the resistor matching.
yeah you could use the negative portions, but the problem would be the same because the angle would still be out of whack!
@@MoritzKlein0 oh yes, sure now I see it haha. The "shift" would be at the other end of the cycle and would probably sound exactly the same...
I'm excited to try this - I've built the VCO up to the point of adding the pitch potentiometer but I dont have a sequencer so I'm looking into writing a simple sequence for my Arduino. If anything I can sample the output and use it in a sampler.
How do you create a sawtooth wave? Help :(
I noticed on the scope trace at the peak of the triangle that there was a line going down to zero. I assumed that is was a kind of crossover distortion caused by the diodes both being in their non conducting region (less that 0.6). Would it be beneficial to use Schottky diodes instead of silicon?
11:31 is a high pass filter the same thing as AC coupling?
yes, just with a low cutoff frequency
would this be the same as a rectifier?
Thanks brother, helpful as always. Great to hear you swear too ;)
ha i was actually hesitant to put that in the script but it just had a nice ring to it!
Ever thought about quartz buffer amp? Would that be usefull for oscilopes or..like a more precise aparatus
I was wondering, how come the forward voltage drop of the diodes does not cause the cutting off of the square wave when Vout from the inverting buffer is less than the Vforward of the diode?
yeah i skimmed out on the explanation there. it works because the 20k resistor after the two diodes is connected to -12V (instead of ground). so even when the op amps are at 0V output, the diodes stay forward biased.
Oh yea that makes sense, didnt notice the -12 V there:)
Don't you get a twice smaller amplitude on the output?
Awesome videos! I was so inspired i'm doing the first part vco right now.
Great video! Was just wondering about the triangle's output voltage range... would that be half of the saw's voltage range?
yeah! if you want to get a 10Vpp output, you could either add another op amp at the end, or you could turn the first buffer into an amplifier with a gain of 2.
Really cool! I wonder if you could've used a integrator, taking a square wave to a triangle directly
correct me if i'm wrong, but i think integrators have to be tuned for the exact frequency of the oscillation. meaning that you can use them to convert a 500 Hz square into a nice triangle, but as soon as that frequency changes, the triangle-shape deteriorates heavily.
@@MoritzKlein0 With what I remember from college, there shouldn't be any weird frequency relations, but there might be a practical reason I'm not aware of
@@ggyshay7687 hmm i'll have to look into that!
@@ggyshay7687 ah sorry, i was wrong. it's actually the amplitude that decreases (pretty massively) as the input frequency increases. you can try it out here: tinyurl.com/yxgkt3vy
@@MoritzKlein0 uhmm the decrease seems really bad indeed
I do not understand how the diode drops disappear?
I expected you to do the neg feedback after the top diode and use another op-amp in a +1 configuration to get rid of the bottom diode drop (again with the feedback after the diode)
This would also make the group delay between the up and down parts of the wave the same.
I have not checked if all that is correct, but my main point is where did the diode drops go?
....or am I missing something?
read the description, i skimmed on the explanation there - sorry!
I'm assuming there's a reason to use this kind of a circuit instead of something more like a bridge rectifier... I just don't know what that reason is. 😆
Yeah I was thinking the same.
Another fantastic video. Nice approach to quick and simple tutorials too. If I may ask, where did you learn about all this stuff? You explain it all very well, meaning you have a great knowledge on the subject. Are you self taught? Or did you take lessons / have a mentor? Thanks again.
thanks! i'm self-taught, mostly learned through trial & error on the breadboard. since i don't have a background in EE, i found most tutorials really difficult to follow and so i wondered if there wasn't a more accessible way of teaching this stuff. which lead me to starting this channel!
@@MoritzKlein0 I am glad you started. This is by far the best channel on the subject and you do it well. Keep it up! I am also self taught, dove right into the deep end 4 years ago wanting to clone the TR-808 sounds and now working on a 32-step sequencer without using a microcontroller. It is quite a challenge, since I want it to do all kinds of crazy things, but it is great fun.
I will eventually make a groovebox combining a sequencer and all of your ideas. I can imagine something like a Monotribe on steroids haha!
Excellent. Thanks !
Hi all. may I ask for your help? I have build Moritz Klein's VCO and with success. I added also the saw to tri converter on the same board too and with the wave mixing (blending) option, I can blend between Saw and square to a single output jack. so I had this idea to put a SPDT switch into the circuit and do the same thing with the triangle and sine wave. Now i have this 1 thing that bugs me a bit. I noticed on my oscilloscope that the triangle wave is not as loud as the other three waves. How can I pump it up a bit so it matched the other three? these are nicely 5v.pp. Triangle sits around 2v.pp. thanks in advance for any help or guidance.:)
try this: tinyurl.com/yfcyfe27 - this also fixes the triangle's "buzzyness". be aware that you need to use a precision trimmer to set the waveshape though.
I love these videos! So clear and understandable! Out of curiosity, is there a reason you didn’t put the diodes in a precision rectifier configuration? It’s the same parts, but uses the op-amp to mitigate the diode drop.
mostly because i was just following mfos/ray wilson‘s lead (musicfromouterspace.com/analogsynth_new/SAWTOTRICONVERT/SAWTOTRICONVERT.php) - interesting concept though, hadn’t heard of it yet!
@@MoritzKlein0 Ray Wilson = Legend. Can't go wrong there!
Great Video! thanks!
Is the Inverting Buffer considered also Inverting Amplifier with a Gain of 1? The way you call it. Kinda makes more sense calling it a inverting buffer, because it is Designed to only output 1 to 1 Inverted Signal.
yes, you can also call it a 1:1 inverting amp!
Hola... buenisimo video..!!!
Hay forma de hacerlo funcionar con fuente simple..??? Si es posible con 9v...
No importa si el triangulo no queda perfecto....
Gracias...!!!
Great video, but isn't the triangle wave the half frequency of the sawtooth? So if we switch the sawtooth from our VCO to triangle we suddenly got big drop in the tone of the sound it produces.
I think the frequency is the same. If you look back at the drawing of the overlapping waveforms, the peaks of the input sawtooth wave and output triangle are at the same positions.
WOW!!!
THANK
YOU
Very nice explanation, you are such talented teacher, i've been building the entire VCO serie circuit, it is on sweet spot.
One question, why did you use TL074 for this case? Can a TL072 be enough?
I'm trying to boarding all in the same pcb but 3xDIP14 chips are quite large.
Thanks for the big legacy.
no reason - i think i just didn’t have anything else on hand. so yeah, using a TL072 will work just fine!
Hi @Moritz Klein and thanks for your instant reply, I really appreciated it !
I've ended just today the layout on a veroboard 21x21 based on TL072 for the SAW-to-TRI, and including the latest of your schematic for PWM and SAW.
When I'll end the build I may post something!
Till then, thanks a lot for your brilliant work!
Dear Klein, mission accomplished!
Your VCO works brilliantly combined with a baby-sequencer, I've just finished the vero-layout and tested it out.
It sounds amazing, even if I'm not even closer to synth professional
Thanks ! Keep it up!
You just "divide" your saw out in two, one goes to the saw jack and the other goes to the converter circuit, right??
Great video as always!!
Thanks a lot mate!! Haste mal wieder reingehauen ;)
top notch. Thanks.
Hi! I'm in love with your videos, they are super clear and very useful at the same time.
However, I have a question about AC coupling.
I don't think I really got why we need the resistor going to ground. Especially because I've seen guitar pedals without that
without the resistor, the wave will not be centered around zero volts. think of the resistor like an equalizing valve; i explained this in detail here: th-cam.com/video/QBatvo8bCa4/w-d-xo.html
you can check out a simulated example here: tinyurl.com/y42zvbx2 - note how the input sine wave has a 5V DC offset, and the output has the same offset. add a 200k resistor to ground though, and your wave is properly centered: tinyurl.com/y4thaukm
i think (correct me if i'm wrong) the reason why some circuits omit the resistor is because they assume the next device will have a resistor to ground at the input terminal.
@@MoritzKlein0 I don't know, that's why I didn't get it...
Particularly, I was looking at the schematic here, which doesn't have a resistor to ground at the input: www.wamplerpedals.com/blog/uncategorized/2020/05/how-to-design-a-basic-overdrive-pedal-circuit/
I don't know if either the tone or the volume control solve this, because they are in fact going to ground (especially the volume).
Thanks for letting me discover falstad.com, though, I have the feeling I will find it extremely useful
14:10 That was mesmerizing lol
So.. I've been meaning to ask. Do you have any recommendations for op-amp and diode selections to build some of your circuits? Is it the case where you have a couple different types of each and use them for most projects or do you generally end up special ordering for most projects?
he generally put the components used, but if you are having trouble finding some of those, you can always look at the datasheets from said components, and from that data find similar components that can be used as replacements
@@Tsukiumy thanks for letting me know!
This is the perfect companion to Ray Wilson's (RIP) saw to triangle converter circuit over at musicfromouterspace.com. Thank you!
Looks like that's where the circuit came from. musicfromouterspace.com/index.php?MAINTAB=SYNTHDIY&VPW=1745&VPH=791
Hi Moritz! Great videos!!! I love your channel and will be suscribing to your patreon. I built this VCO from scratch and I noticed that, when applying a CV signal, the waveforms at the output increase their gain from 10vpp to 12vpp aprox... is this normal? I've never been aware of this before.
Again, kudos on your videos!!! Great work!!! Thank you.
love it! thank you!
Awesome videos! I've been using them to make my own synths :) I was just curious why you would use an inverting op amp instead of a full bridge rectifier? Is there an advantage?
Can you explain which SAW signal is the input to this circuit? Based on the VCO schematic, is it the TP3 (Test point 3)? Or is it XS4 (The output to the SAW wave jack). It's not clear to me. Love the series, and it has inspired me to get into DIY :D
So far I'm assuming it is TP3. The provided simulation in the description shows the 1uF + 100k which is identical to the TP3 point. I think seeing the 200k value in the video threw me off
Honestly, I find that the reverse is more useful. I have way more use for a saw wave than a triangle. But if you need a triangle wav, use a square wav to feed the integrator. Saw waves are precious. That's why the "super saw" exists and not the super triangle or super square.
the problem is if you run a square through an integrator, the waveform morphs if you change the frequency.
@@MoritzKlein0 In an LFO or even say... 1Hz-20Khz? I was pretty sure you can get a stable triangle in the narrow audio frequency from a square. I'm gonna have to do some experiments.
thanks!
great!!
> quick tip
> 15 minutes
Hell yeah
everything is relative!
The circuit you need is a Precision Full-wave rectifier; yours is not quite right but it only needs two tiny changes.
First, put the lower diode where the feedback resistor is, maintaining the polarity. Then attach the feedback resistor to where the two diodes currently join. Feed the output into a buffer amp as you suggest, but instead of AC coupling it, mix in the original sawtooth as well to level shift the output